(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Bulletin of the British Museum (Natural History), Mineralogy"

SOME ASPECTS OF FENITIZATION 

WITH PARTICULAR REFERENCE 

TO CHILWA ISLAND AND 

KANGANKUNDE, MALAWI 



A. R. WOOLLEY 



BULLETIN OF 
THE BRITISH MUSEUM (NATURAL HISTORY) 
MINERALOGY Vol. 2 No. 4 

LONDON: 1969 



SOME ASPECTS OF FENITIZATION WITH 

PARTICULAR REFERENCE TO CHILWA 

ISLAND AND KANGANKUNDE, 

MALAWI 



BY 

ALAN ROBERT WOOLLEY 

British Museum (Natural History) 



Pp. 189-219; 12 Text-figures, plate 9 



BULLETIN OF 
THE BRITISH MUSEUM (NATURAL HISTORY) 
MINERALOGY Vol. 2 No. 4 

LONDON: 1969 



THE BULLETIN OF THE BRITISH MUSEUM 

(NATURAL history), instituted in 1949, is 
issued in five series corresponding to the Departments 
of the Museum, and an Historical series. 

Parts will appear at irregular intervals as they become 
ready. Volumes will contain about three or four 
hundred pages, and will not necessarily be completed 
within one calendar year. 

In 1965 a separate supplementary series of longer 
papers was instituted, numbered serially for each 
Department. 

This paper is Vol. 2, No. 4 of the Mineralogy 
series. The abbreviated titles of periodicals cited 
follow those of the World List of Scientific Periodicals. 



World List abbreviation 
Bull. Br. Mus. nat. Hist. (Miner.). 



Trustees of the British Museum (Natural History) 1969 



TRUSTEES OF 
THE BRITISH MUSEUM (NATURAL HISTORY) 

Issued 14 October, 1969 Price £1 



SOME ASPECTS OF FENITIZATION WITH 

PARTICULAR REFERENCE TO CHILWA 

ISLAND AND KANGANKUNDE, 

MALAWI 

By A. R. WOOLLEY 

SYNOPSIS 

A general penological account is given of the process of fenitization at the carbonatite com- 
plexes of Chilwa Island and Kangankunde. Twelve whole rock analyses are presented and the 
chemical changes occurring during fenitization are discussed in the light of these new data, 
together with published data from other complexes. Particular emphasis is placed on the use 
of the triangular diagram quartz-nepheline-kalsilite for following the changes in the felsic 
constituents during fenitization of granitic rocks, and for displaying the chemical differences 
between the processes of fenitization, potash-feldspathization, and nephelinization. The 
possible relationship between these processes and carbonatite emplacement is discussed, and it is 
suggested that there is probably an evolutionary series of carbonatites having strongly contrasting 
metasomatizing powers. It is pointed out that most fenite occurrences can be linked with 
silicate magmas, and that fenites which are undoubtedly attributable to the action of carbonatite 
are rare. 

INTRODUCTION 

The Chilwa Province of Southern Malawi comprises a series of volcanic vents, which 
are filled with carbonatite, feldspathic breccia, agglomerate and nepheline syenite, 
plugs of syenite and nepheline syenite, and an associated dyke suite of solvsbergite, 
trachyte, microfoyaite, phonolite and nephelinite. (Garson 1965&, p. 12) (Text-fig. 1.) 
The province appears to be Upper Jurassic to Lower Cretaceous in age (Bloomfield, 
1965), and is associated, in a broad sense, with the rift system. These rocks were 
first described in the classic bulletin of Dixey, Campbell Smith & Bisset (1937), 
and they recognized the intrusive nature of the limestone on Chilwa Island, the 
first carbonatite to be described from Africa. The Malawi carbonatites have, in 
recent years, been investigated and described in great detail by Garson (1962, 1965a, 
19656; and numerous other papers) and Garson & Campbell Smith (1958). The 
Chilwa Island and Kangankunde vents are well suited to a study of fenitization 
because of the broad extent of their aureoles, but are restricting in so far as the outer 
limits of the Chilwa aureole are concealed beneath the lake, while the outer zones of 
fenitization at Kangankunde are poorly exposed. At both localities a group of 
potash-rich rocks, mapped as feldspathic breccia and contact breccia by Garson at 
Chilwa Island (1958, Map 2) and as feldspathic breccia and agglomerate at Kangan- 
kunde (1965a, Map 2), lie between the carbonatites and the fenites, and these rocks 
are characteristic of many of the Chilwa Province carbonatite complexes. They have 
been shown to be derived from the more normal fenites by a process of potash 
metasomatism (Garson & Campbell Smith, 1958, p. 30; Garson, 1965a, p. 37). These 

MINER. 2, 4. 13 



192 



FENITIZATION AT CHILWA ISLAND 



potash-rich rocks, as well as the more typical fenites, are included in this study. At 
neither locality do extensive bodies of intrusive silicate rocks occur, although small 
plugs and dykes of nepheline syenite, ijolite, trachyte, nephelinite, solvsbergite and 
alnoite are found at Chilwa Island (Garson & Campbell Smith, 1958) and a few dykes 
of solvsbergite and alnoite have been emplaced at Kangankunde (Garson, 1965a). 

The fenitization at both these localities has already been described (op. cit. 1958, 
p. 28) so that the field and petrographic data presented in this account will be of a 
general nature and a stress will be put on new observations. As well as over 300 
specimens collected by me, use has been made of the thin-sections in the collection 
of the British Museum (Natural History), made for Dr. Campbell Smith from 



LAKE 
CHILWAl 



• jTundulu 




agglomerate vents 
fc&'fj Nepheline syenites 



.V.*! Syenites & granites 
I J Lupata volcanics 
10 20 30 



Imiles 



Fig. 1. Distribution of the rocks of the Chilwa Alkaline Province in Southern Malawi. 
Map based in part on map of Garson (1966, fig. 1). 



AND KANGANKUNDE, MALAWI 



193 



material collected by Dixey & Garson. No petrochemical data of the fenites of 
Chilwa Island or Kangankunde have been published heretofore. 



PETROLOGY 

Chilwa Island 

Geologically, Chilwa Island is built of a central core of carbonatite which is 
surrounded by a collar of feldspathic breccia and this in turn is surrounded by a group 
of fenitized schists and syenites. A simplified geological map, taken from Garson 
& Campbell Smith (1958) is given as Text-fig. 2. Garson has distinguished and 
mapped three types of carbonatite, while he divides the fenitized basement into a 
group of granulites and a series of older intrusive syenites (op. cit. 1958, map 2). 
The granulites vary from quartzo-feldspathic schists through to pyroxene and 
hornblende-quartz-feldspar types, and limestones, amphibolites, pyroxene schists 
and rocks of charnockitic affinities also occur. The syenites are coarse rocks, usually 
of a porphyritic habit. 

The fenitized rocks are characteristically net veined and display a greenish hue, 
features described from many fenite localities, while the feldspathic breccias are red 
or cream in colour and distinctly more leucocratic. No contacts between the fenite 
and the feldspathic breccia can be seen. From a study of the thin sections it is 
possible to distinguish two main types among the fenites, a less fenitized group of 
quartz fenites and an inner group of syenitic fenites. The distribution of the two 
types is brought out in Text-fig. 3A. The quartz fenites are distinguished in thin 
section by a mineralogy inherited primarily from the basement including plagioclase 
(oligoclase to andesine), orthoclase, hornblende, ortho- and clino-pyroxene, biotite 
and quartz. The syenitic fenites, in contrast, are dominated by a new mineralogy 
of perthite, aegirine, minor alkaline amphibole and secondary biotite; quartz is now 
rare. The feldspathic breccias are formed dominantly of potash-feldspar and ore, 
with minor zircon, although commonly they contain quartz, the result of late 
silicification. The range of minerals in the fenites is as follows: 



Country 
rocks 



PRIMA R Y MI N ERA LS 
Hornblende, ortho, clino-pyroxene, 

ore and biotite 
Plagioclase and orthoclase 
Quartz 

SECONDARY MINERALS 
Aegirine 

Riebeckitic amphibole 
Magnesioarfvedsonitic amphibole 
Secondary biotite 
Ore 

Perthite 
Orthoclase 
Quartz 



Quartz 
fenites 



Syenitic 
fenites 



Feldspathic 
breccia 



194 



FENITIZATION AT CHILWA ISLAND 




■ 



Carbonat ites 

Feldspath ic 
breccia 



Fenites 

Superficial 
deposits 



Locations of 
analysed specimens 



1 mile 



Fig. 2. The geology of Chilwa Island. Based on map of Garson & Campbell Smith. (1958, fig. 9). 
The locations of the analysed specimens are indicated. 



AND KANGANKUNDE, MALAWI 195 

The change from the quartz fenites to the syenite fenites is vigorously promoted 
through the microscopic network of cracks, which develops early in the fenitization 
process, and along the length of the fine network of "fenite " veins which later form 
along these cracks. Quite quickly the rock along, and immediately adjacent to, 
the veins, by the growth of aegirine and alteration of the feldspar, becomes a syenitic 
fenite, and by the pervasion of these changes through the whole rock the syenitic 
fenite stage is reached. The alteration of the syenitic fenites to the feldspathic 
breccias is also promoted through the agency of the fenite veins in so far as the 
alteration of the aegirine, amphibole, and biotite to ore minerals first takes place 
along the veins, and eventually the rock is reduced to potash feldspar with a fine 
vein system defined by the ore which pseudomorphs the characteristic veins of the 
syenitic fenites. 

The sequence of mineralogical changes which have been observed to take place in 
the alteration of the basement schists and syenites into the syenitic fenites are as 
follows : 

PRIMARY MINERALS SECONDARY MINERALS 

Hornblende > aegirine ± magnetite ± biotite ± alkaline 

amphibole 

Hornblende > biotite (magnetite) — ■ > aegirine 

Biotite > aegirine i alkaline amphibole 

Ore > alkaline amphibole 

Clino-pyroxene > aegirine 

Ortho-pyroxene > riebeckitic amphibole 

Quartz > aegirine 

Plag. and orthoclase > perthite ± plagioclase i microcline 

The alteration of hornblende to aegirine often involves an intermediate stage of 
biotite, and sometimes magnetite, growth and in the case of biotite a very thin rim 
of an unidentified feldspar usually intervenes between the mica and the replacing 
aegirine. The secondary biotite is itself replaced by aegirine, but it persists through- 
out the syenitic fenite zone. It is pale to medium brown in colour and probably 
has a high ferric to ferrous iron ratio as reported from Alno (von Eckermann, 1948, 
p. 32). Aegirine is the most widespread and characteristic new mafic mineral in the 
fenites. It develops at the expense of hornblende, biotite and pyroxene and replaces 
quartz, often by the formation of a peripheral corona of radiating prisms which 
gradually encroach inwards. Aegirine also grows as isolated grains and decussate 
and radiating clusters along the length of the " fenite " veins; indeed these veins are 
commonly built solely of aegirine. With increasing fenitization the aegirine aggre- 
gates recrystallize to form much larger individual grains. The colour of the aegirine 
varies from a deep blue-green to deep emerald-green through to buff or almost 
colourless, and radiating clusters are often colour zoned, invariably with the colour 
intensity increasing outwards. Optical properties do not appear to change with 
colour. An aegirine-augite with a : c = 20 has been reported (Garson & Campbell 
Smith, 1958, p. 21). 



ig6 



FENITIZATION AT CHILWA ISLAND 





Fig. 3. (A) The distribution of syenitic fenites, quartz fenites, and lightly fenitized rocks 
on Chilwa Island, based on thin-section evidence. ■ = syenitic fenites; B — quartz 
fenites; □ = lightly fenitized rocks. Squared ruling = carbonatite. Dots = felds- 
pathic breccia. (B) Distribution of alkaline amphiboles in Chilwa Island fenites. 
• = pale blue amphiboles only (magnesioarfvedsonite) ; C = mixed pale blue and 
riebeckitic amphiboles; O = riebeckitic amphiboles only. 



AND KANGANKUNDE, MALAWI 197 

Two distinct types of alkaline amphibole occur. One is a deep blue riebeckitic 
type which is confined to the outer fenite zone, particularly along the south coast of 
the island and pyroxene-bearing rocks on Chaone peninsula and at Marongwe and 
Chirunda Hills, in which it forms at the expense of ortho-pyroxene (Text-fig. 3B). 
The second amphibole is characteristically of a pale lilac colour and is almost univer- 
sally present, though in small amounts, among clusters of aegirine grains in the 
syenitic fenites, and the inner zones of the quartz fenites; it is probably a magnesio- 
arfvedsonite (Deer et al., 1962, p. 364). It is not certain whether there is a change in 
composition of the amphiboles with increasing fenitization in line with an increase 
of the oxidation ratio during fenitization, or whether the amphibole paragenesis is 
dictated by original rock composition. The optical properties of these amphiboles 
are difficult to determine because of the small size of the grains, deep colour, strong 
absorption or anomalous extinction. The riebeckite is pleochroic in shades of lilac, 
blue and buff e.g. a = deep blue; /? = pale yellow-buff; y = indigo-blue; a. : c = 3 . 
The paler amphibole has already been discussed at length (Dixey et al., 1937, p. 18; 
Garson & Campbell Smith, 1958, p. 21). However, in the latter account, the extinc- 
tion is described as small, but during the present work angles of a : c = 42-44 
have been measured on many grains. An amphibole having a strength of colour 
somewhere between these two amphiboles and a : c = 30-35 ° sometimes occurs in 
the quartz fenites. The relative distribution of the amphiboles is shown in Text-fig. 

3B. 

Plagioclase, oligoclase to andesine, is the dominant feldspar in the basement schists 
and is important in the syenites. During fenitization it is transformed into a perthite 
as is the orthoclase. The course of this alteration is often well illustrated along the 
fine veins and cracks which traverse the feldspar. Along and adjacent to these veins 
a zone of turbidity develops, which gradually intensifies and spreads into the feld- 
spar, while concomitantly the plagioclase twinning fades. The turbidity also spreads 
in from the feldspar margins, indicating that the crystal boundaries also were 
channels for fluid migration. Textural evidence suggests that the formation of the 
turbidity is probably associated with the removal of sodium from the feldspar, 
because several examples have been observed of narrow veins crossing feldspar, in 
which they cause turbidity, then, in crossing adjacent quartz grains promoting the 
growth of aegirine crystals. It appears likely that the sodium fixed in the aegirine 
has been transported along the vein from the feldspar. Narrow veins of carbonate 
have also been observed to generate turbid aureoles in plagioclase and to destroy 
the twinning. Orthoclase similarly becomes turbid early on in the fenitization. The 
syenitic fenites are dominated by a feldspar phase which often shows typical exsolu- 
tion perthite textures, similar to those described from other fenite localities, e.g. Fen 
(Saether, 1957, fig. 4), and Alno (von Eckermann, 1948, Plate 10, fig. 1). In the 
syenitic fenites formed from the porphyritic basement syenites the forms of the 
original feldspars are picked out by the turbidity, these turbid relics usually being 
surrounded by fresh rims of alkali feldspar which may be finely twinned. Some quite 
fresh albitic plagioclase and twinned microcline occurs in the syenitic fenites in the 
north-east corner of the Island near to Kotamu as a fine-grained granular mosaic 
insterstitial to the larger feldspars. This fresh, new feldspar appears to be 

MINER. 2, 4. I3§ 



i 9 8 FENITIZATION AT CHILWA ISLAND 

associated, in general, with thick fenite veins of aegirine and potash-rich feldspar, 
but it could represent recrystallized material exsolved from the larger feldspars. 

The mineralogical changes involved in the transformation of the syenitic fenites 
into the feldspathic breccias are relatively straightforward. The aegirine becomes 
unstable and breaks down to ore which, in contrast to the magnetite of the quartz 
and syenitic fenites, appears to be a mixture of haematite and limonite, while the 
feldspar, in which Carlsbad twinning is widespread and microcline cross-hatching 
occasionally to be seen, loses the perthite structures and becomes very much more 
turbid. Zircon is now quite abundant and quartz returns in the form of drusy 
infillings, which tend to be concentrated along lines of brecciation and narrow veins 
cutting the feldspars. Some varieties consist almost exclusively of potash-feldspar. 
The feldspathic breccias represent the ultimate stage of fenitization at Chilwa 
Island in so far as they lie adjacent to the carbonatite and are included as blocks 
within it. 

The veins which cut the fenites are of two types: carbonate veins, which are 
probably direct emanations from the carbonatites; and the fenite veins proper. 
Within the carbonate veins phlogopite, barytes and anatase have been identified. 
The fenite veins vary continuously from hair-like cracks and lines of turbidity 
within the feldspars of the outermost fenites, via veins of aegirine as much as 0-5 cm. 
wide, to aegirine-potassium-feldspar veins up to 10 cm. thick in the syenitic fenites. 
The aegirine veins (Garson & Campbell Smith, 1958, fig. 3) cut cleanly across the 
rock-forming minerals but the narrower ones are often discontinuous, fading out and 
starting again quite haphazardly. As well as aegirine lesser amounts of magnetite, 
which is occasionally the sole vein mineral, and riebeckite in the quartz fenites and 
magnesioarfvedsonite in the syenitic fenites, are found. The coarser veins are 
built of selvages of acicular, deep-green aegirines and cores of microcline-micro- 
perthite. The growth of the aegirine perpendicular to the walls gives rise to a comb- 
structure. Carbonate, magnetite, quartz and bastnaesite may also be present. 
The remnants of the veins can be identified in the feldspathic breccias as linear 
concentrations of iron oxides. 

Kangankunde 

The fenitized rocks of Kangankunde have been sub-divided, for petrographic 
purposes, by Garson & Campbell Smith (1965a, p. 29) into an outer group of shock- 
zone fenites, and an inner group of permeation fenites. Adjacent to the main 
carbonatite centres the fenites have been mapped by Garson as " feldspathized 
fenites ", between which and the carbonatite " feldspathic breccia and agglomerate " 
has been distinguished (Text-fig. 4). The feldspathic breccia and agglomerate is 
equivalent to the feldspathic breccia and contact breccia at Chilwa Island. The 
process of feldspathization has been taken a stage further at Kangankunde however, 
by a process of phlogopitization (op. cit. p. 42), which is not found at Chilwa Island. 

The fenitized rocks of Kangankunde show a number of marked differences from 
the fenites of Chilwa Island: 

(1) A wider range of basement rocks come within the Kangankunde aureole, 



AND KANGANKUNDE, MALAWI 



199 



including amphibolites, garnet and epidote amphibolites, meta-dolerites, hornblende- 
biotite gneisses, quartzo-feldspathic granulites, quartz reefs and granitic pegmatites, 
while the syenites, which occupy about half of the basement at Chilwa Island, are 
absent here. 

(2) The lower grades of fenitization are fully represented at Kangankunde. At 
Chilwa Island this zone is mostly hidden beneath the lake. 



< 



*'.on V 



■^ 



Carbonatite and 
carbonatised rocks 

Feldspathic breccia and 
feldspathised fenite 



Locations of analysed 
specimens 



FEET 
500 1000 1500 




Southern 
. Knoll /; 



Fig. 4. Simplified geological map of Kangankunde based on map of Garson (1966, fig. 7). 
The locations of the analysed specimens are indicated. 



200 FENITIZATION AT CHILWA ISLAND 

(3) Fenitization proper (see p. 209) is not so intense at Kangankunde, as evidenced 
by the restricted occurrence of syenitic fenites. 

(4) The oxidation of the ore minerals and the breakdown of aegirine and alkaline 
amphibole to hydrated iron oxides, which are characteristic of the feldspathic 
breccias of Chilwa Island, are more widespread at Kangankunde, and are often 
manifest well outside the zone of feldspathized fenites, as mapped by Garson. 

(5) Primary quartz persists well into the feldspathic breccia and agglomerate. 

(6) Alkaline amphiboles are more abundant in the Kangankunde than the Chilwa 
Island fenites. They appear to represent a greater range of compositions and have 
quite different textures to those of Chilwa Island. 

(7) Carbonate is widespread. Some of the carbonate has resulted from the altera- 
tion of plagioclase, but most of it is concentrated along the centre of, and adjacent 
to, fenite veins, and a close correlation can often be demonstrated between these 
carbonate-bearing veins and the intensity of fenitization of the enclosing rock. 
Carbonate is usually only evident in the inner zones of fenitization at Chilwa 
Island. 

(8) The coarse aegirine-potash-feldspar veins found at Chilwa Island have not been 
found at Kangankunde. 

(9) Many of the feldspathic breccias of Kangankunde have been phlogopitized. 

Many of the detailed mineralogical changes revealed by thin-sections of Kangan- 
kunde fenites are similar to those described for Chilwa Island. The basement 
plagioclases (oligoclase to andesine) are made over to an intensely turbid potash 
feldspar along veins and cracks and around the margins but the development of 
perthite textures, which characterize the syenitic fenite stage, has only been observed 
in a few specimens from the South-east side of the Fenite Spur (Garson, 1965a, 
Map No. 2). The primary hornblende, epidote, biotite, and garnet of the basement 
gneisses are gradually replaced by feldspar and altered to aegirine, alkaline amphibole 
and biotite. Whereas hornblende usually alters directly to aegirine in most Chilwa 
Island fenites, an intermediate fine-grained biotite stage is usual at Kangankunde. 
The primary biotite commonly persists into the feldspathic breccia stage. Alkaline 
amphiboles tend to be intergrown with aegirine and form fine-grained, fibrous, 
stellate, zoned masses in which the zoning may be defined by colour changes or by 
alternate layers of amphibole and aegirine. These masses replace quartz and mafic 
minerals and also form nodules and larger masses along fenite veins. Garson 
(1965a, p. 33) was able to distinguish, by optical means, riebeckite, crossite, 
magnesioarfvedsonite and soda-tremolite, often intergrown in a single mass (op. cit. 
fig. 3). The abundance of iron oxides, apparently in the form of limonite and 
goethite, is characteristic of the Kangankunde fenites, and in the higher zones of 
fenitization these completely replace the mafic minerals and extend as a coating to 
the feldspars. The breakdown of the mafic minerals is often strikingly promoted in 
the vicinity of carbonate concentrations, and carbonate is invariably abundant 
where phlogopitization of the potash-feldspar has taken place. The phlogopite 
development is most widespread in the feldspathic breccia and agglomerate and 
within xenoliths of these rocks in the carbonatite, but it can be observed also in the 
more carbonate-rich of the feldspathized fenites. 



AND KANGANKUNDE, MALAWI 201 

It is everywhere apparent in the Kangankunde fenites that the process of fenitiza- 
tion has been effected mainly through the agency of the fenite veins, although the 
extreme development of these veins, as represented by the thick aegirine-potash- 
feldspar veins of Chilwa Island, is not found at Kangankunde. The chemical 
changes which are associated with the early development of the fenite veins have been 
investigated with an electron probe. Some of the results are given as Plate 9. 
The photomicrographs show the way in which the multiple twinning of the plagio- 
clase is destroyed along the veins and is replaced by a turbid, untwinned feldspar. 
Along the centre of the veins there is commonly a thread of aegirine. The probe 
traverses indicate that the turbid feldspar is strongly potassic, sodium being con- 
centrated in the aegirine and plagioclase. The probe work also showed that the 
potash-feldspar is very much lower in calcium than the plagioclase. This alteration 
represents, on a small scale, the development of a syenitic fenite. The same process, 
on a hand specimen scale, is shown in Plate 9 B and C in which the altered feldspar 
zone is now a centimetre or more wide. This contrast constitutes the junction 
between the fenites and the feldspathized fenites, and for Chilwa Island is equivalent 
to the hidden contact between the feldspathic breccia and the syenitic fenites. Thin- 
sections across the fenite vein of Plate 9 B reveal that the inner part of the vein is 
built of turbid feldspar and ore — equivalent to the potash-rich feldspathic breccias; 
the outer parts of the vein contain abundant pyroxene, amphibole and turbid feld- 
spar — equivalent to the syenitic feldspar stage, while the unaffected rock is a quartz 
fenite. 



THE CHEMISTRY OF FENITIZATION 

In Table 1 are given the results of chemical analysis of 12 fenites and associated 
rocks from Chilwa Island and Kangankunde, the localities of which are shown on 
Text-figs. 2 and 4. When considering bulk chemical changes during fenitization it 
is probably advantageous to compare cationic concentrations on a standard oxygen 
cell. The case for this procedure has been argued fully by McKie (1966, p. 262) 
and following McKie the cationic concentrations per 100 oxygens are given in Table 2, 
and used for comparative purposes. 

Of the Chilwa rocks, six are fenitized older syenites, one is a thick " fenite " vein 
and one a feldspathic breccia. The analysed fenitized syenites are now syenitic 
fenites composed dominantly of aegirine and perthite, with minor quartz, alkaline 
amphibole, carbonate, ore, secondary biotite and plagioclase, and quartz fenites 
containing appreciable quartz, and lesser amounts of hornblende and biotite, and 
in one specimen (No. 4) clinopyroxene. A suite of older syenites were chosen for 
analysis because these rocks are more homogeneous, prior to fenitization, than 
the very variable basement schists. The analysed fenite vein is 6 cm. thick and has 
selvages of deep-green, acicular aegirine forming radiating clusters, dominantly 
orientated perpendicular to the vein walls, and an interior composed of stubby, 
subhedral perthite plates up to 1 cm. across. Minor ore and sericite are present. 
The cream coloured feldspathic breccia consists of about 98 % potash-feldspar which 
is extremely turbid, as seen in thin section, and forms a granular mosaic and in 



202 FENITIZATION AT CHILWA ISLAND 

places a vaguely denned trachytic texture of subprismatic, twinned grains. Ore and 
zircon are also present. 

Garson (1965a, p. 124) has pointed out that because of the wide range of rock 
types in the Kangankunde aureole it is not practicable to obtain a series of chemical 
analyses of rocks to show the trend of fenitization. Therefore the two veined 
specimens illustrated in Plate 9 have been chosen as representing, on a small scale, 
the changes from the quartz fenite through to the syenitic fenite and feldspathic 
breccia stage. The main fenite rock and the veined parts were mechanically 
separated and both portions analysed. The specimens were found in dry stream 
beds on the south side of Kangankunde, west of the Southern Knoll (Garson 1965a, 
map 2). Neither specimen was in situ, but both probably come from the rocks 
mapped as " feldspathized fenites " by Garson, which occupy the higher slopes of the 
hill. 

Rock B.M. 1968, P 37, 284 (analyses 9 and 10) is essentially a leucocratic quartz-feldspar 
granulite. Quartz comprises about 25% of the rock and the feldspar is mainly a sodic plagio- 
clase, with some potash-feldspar. The feldspars are sericitized around the margins and along 
cleavages and cracks. A reddish biotite and magnetite are plentiful. The metasomatic fenite 
vein comprises a central thread of limonite or goethite, and an aureole surrounding this in which 
the feldspar is much more turbid than in the matrix fenite (Plate 9 B) and there is a great abun- 
dance of colourless to pale-green stellate clusters of aegirine needles. Quartz decreases within 
this aureole, and biotite is rarely found. The changes in the feldspar are vividly shown in places 
in which twinned plagioclase grains are sharply truncated by the veins, as previously described 
(Plate 9 A). The significant changes associated with the veining are, therefore, the decrease of 
quartz, development of aegirine, oxidation of the ore minerals and the change of the plagioclase 
to a potash feldspar. This overall transformation is consistent with the changes involved in 
going from a quartz fenite towards a syenitic fenite. 

The changes along the vein of rock B.M. 1968, P 37, 299 (analyses 11 and 12) are even more 
profound, in so far as they appear to represent the transition from quartz fenite, through syenitic 
fenite to a feldspar-ore rock of feldspathic breccia type. The main rock is a plagioclase 
(oligoclase-andesine)-quartz-hornblende-biotite gneiss with minor garnet and ore, and some 
secondary aegirine and patchily distributed turbidity in the feldspars. The veins have a thin 
central thread of a hydrated iron oxide and a zone on either side, up to 2 cm. across, which is 
quite pink in the hand specimen, and in thin-section is characterized by the extreme turbidity 
of the feldspars. Within this zone there is a radical decrease in the quartz and no hornblende, 
biotite or garnet occur. Aegirine and secondary amphibole are abundantly developed along the 
outer edges of the turbid zone, the inner parts being dominated by feldspar and irregular ore 
patches. It would seem, therefore, that the metasomatic changes which define the presence of 
these veins represent, on a small scale, the variation from quartz fenites, through syenitic 
fenites, towards the extreme type of fenite consisting of potassium-feldspar and ore only. 

The chemical analyses were made by A. J. Easton. Si0 2 , A1 2 3 , CaO and MgO 
were determined gravimetrically on a sodium carbonate fusion, while the other 
oxides were determined on an acid (HF/H 2 S0 4 ) dissolution using spectrophotometric 
and flame photometric methods. 

An inspection of the Chilwa analyses in Tables 1 and 3 reveals immediately the 
extreme chemical differences between the quartz and syenitic fenites and the felds- 
pathic breccia. The breccia is very much higher in K 2 and lower in Na 2 0, and is 
poor in Ti0 2 , Fe 2 3 , FeO and CaO. These changes are obviously to be expected 
from the mineralogy, but the field and microscope evidence suggesting that these 



AND KANGANKUNDE, MALAWI 



203 



W 

m 
< 
H 



M»CO N 
lO 0) O M 



IO N H 10 t-~ ON O 

o in-i-M Tt-o 10 to 



on o 
in 



00 04 



in n o 

H V 



00 o o 00 vo o 



10 

O W M 



O O 



O in 00 
tJ- QN t^ H 



u-> 00 O 
ro CO h in 



01 O 



00 O O CO o o o 



o o 



OOt-fC^IHMOO^OMO 

in co^o 00 m vo vo o< coon 



M N 00 

O 00 in O 01 



O ^1- M o o o 



ON 



O moo 00 in in 00 o ci coco 

t~-MCOMt^NWMU-lMHMU-)>0 



vO O 

vo 



iOhhcohO^NOOON 



cd 

T) 
CI 

9 

bo 

9 
M 

Ft 

Bj 

a 



N M 00 >o in O CO On M 00 

00OONOWO0O "*- m 01 01 in 01 

NOOOOOOOO^OOOO 
vO < 



V 



V 



Tt-N vn n +10 o o< 00 vo 
vo u-jint^M r^iocoM m o m con 



OiOOOtHCOOOTt-COO 



o o 



vO CO On N H O 

o 100 000 



o> >o -*• vo on 

Cl H 01 U-) C~~ H O 






coolOOincoOO 



10 
o o 



c3 

is 
u 



CD 



co 

>. 

"c3 
a 

ni 

"d 
o 



(D 



10 o vo t~» o> o 
o m -^-00 o vo 



+ n o in fi 

On 01 Tt VO lO N 



' 00 M 

in 



+ IOH ■>? O O vO <t O O 



Oi N tJ- Oi O 
M 00 ON M 01 



■*■ co o> in 00 vo 

O On O "tf-oo co m 



co O 
vO 



cOtJ-cn COO "t-Tt-O O 



t^ 01 
O H 



Tj" CO 

b w 



00 r^ c^ w 

CO N "*■ 01 H 



O On 00 

C"» O H 



O O 

l~- "*• H Tj- 00 



NOtNTt-MCOOOoOTt-OOOO 



vO 



or^-oovoooocoinwoi 

o o^coinvo Kid h o t-^-^-M -^-oo 



OIOco^woiOOOO^OOOO 

VO IH 



COinM N lO LH H COMVO CO 

inO\0 h '"dr vO CO 01 vO vO COM 



oi in 



On O 

in 



NvOi-iTj-i-iOC^TfOOOO 



o"0 N °-°"0 O OO 0„ cD 6 6 - 
Q y .-f <d )£> d &p « rt «H, « »o 



On 



ON 

On 



in 

vO 



On 
On 



On 

in 



On 
On 






On 

On 



On 
On 



o 

<N 



O 




o 
in 



On 
On 



On 
00 



ON 
ON 



o 
in 



o 





o 
o 



o 



On 
On 



O 

o 



in 



On 



CO 



o 

o 



O 

ON 



O 

VO 



co 

00 



vO 
vO 



vO 



01 



o 
00 



ON 
ON 



Cu 
00" 

vO 

ON 



m 



I VO 

0> 



CO 

ft 



ft Ph pu, 



00 

VO 

CN 



pq' 



cs 



CO --. 
O 00 ^ 

M O ^f 

. 1-1 o 
r — _ >-< 

»1 N . 
- fO N 

00 -Ph 
vO 00 , 
ON o 00 

M ON VD 
. *-> ON 

vh . 1-1 
^ *3! ■ 

— PQ f- 



CO 

ft ^ 

. ■*• 

CO CO 

vo 01 

ON . 

M t-^ 

. CO 

S Pu 



<D 

T 


s 

Ft 
ni 
M 

Ft 

n) 

MPh 



rO 



pq 



en 



CO 



7> O 



ON 



Ft " 
60 ^ 

o 



li O 



^ 3 ii ,5 






M ON 

d M 



a! 
60 
Ft 

Si 



oSs§ 



o ^ 



In 1^ '^ 

5 



■p o D ai 

'3 ."t; S :fcj 

<u 'c 'c 'c 

>, 0) <U <D 



'S-a 

(U (H, 



£3« 

g -a 

<D Ft 
> _d 
CD CO 



d CD 
O* > 



2 > 
'to 



TJ 



CD 


l-H 


C d 


"o 


CD 


CD fe 


T) 


O 


d p; 


S 


-a 



li ON 

3 3 

CD CD 

N CD ■ 

•J3 ft; 



O^ CD 

•v 6 

CD •- 




Z 

o 

H 

CO 

«i 

w 



'd 
Ft 



d 



3 
CD 



^3 



PQ 

CD 



CD 



MOtfOTt-iovOC^OOONOHM 



CD 



CD 

6 

3 

R 

CD 



3 

CD 

d 

Ph 



u 

CD 

^3 



3 



204 



FENITIZATION AT CHILWA ISLAND 



01 

w 



CO 

CI 

O 

'S 
cS 

Pi 
CD 

bo 

o 

o 

o 

H 

l-l 

ID 
Ph 
CO 

PI 

O 

-.-^ 

ct3 
O 



PI 
CD 
4-1 

CD 

pi 



M 

PI 

ccj 

be 

§ 

W 

C 
g3 

TJ 
Pi 



ctj 

o 

CD 



O 

S 
o 
o 



C^N Tt-O O lO "1 N,? N *00 Tl- I s -. 
^CTii-ihhOwc^Ov 00 ^ 1- '" 1-1 

M roOOcoONOO rr) Tt-0'^-MH 
co m 

oiUN(n(noiOii-!iH tin « r^ 
M t O N CO >o CN iOO f>o o -3- r^ <o 

"OOOOOONOO^OOmOO 
co 

N M CO i- iT)0 C^2}tJ-C>N h m 
— OHOcOMinOOT^ooocoi-iO^ 

m OOc>MOwOO t *"i-iO"-ii-iO 

m 

co o r-^ o O OO u-> J>- oo 00 N co o 
roooo •*!- co O Oi O N i- W t>. co 
O • • " ■ 

OOO^OOmOO~i-iOihOi-i 
ro 

NmNwOOOOO,^L"c>ihOOO 

oo»ni-ioovoo^Ni-ic^r-~M 
oo • • i ■ • • 

ioOcoOOOOO°OOmOO h 

CO M M ^ 

N(OHfltsOOin!f 1 Ot f lOiOiO rt 
lO<N-^-00iO^lOH c ^lN0H0H [-, 

t-» L. 

inOr^"3-ONOO^>OOHOO r 

CO 

<s 

>ot^OM>oioior^^fooiNai<^ -a 

0<OT)-Ot-^vOOOO co 'Omrororo o 
O ' u 

roOMCSHMOO^NONOO 32 

„ H | 

r^-OcomrOTj-i-iro^NOCNr^Th c 

M-iOOO lOiflOOOO H H ^-NiO NOl 

""> • • • |L ■ ■ • - £ 

coOO\CNOcNOO r ^coOcNOO S 

c 

MOOOO'ti-CO+SJflNHCOH S 

i^cocor^Oooc^o~io>-i'ntoo p 

«O0O\HMM00"*"fO0H0l-l R 

co 

CD 

ooronNHioo* 'i-umooco H 

r^ co ^- 00 O O^O O M -^-m lOroO 
CO 1. 

inOOOwwwOO^cOOi-iOO 

CO 

tJ-m WOO CO NC N co O MOO N 

* * O a> N in r^ O o^ tj- n locovo 

CN ■ A • 

ioOOhOmOO^coOi-iOO 
CO 

0\ N CN COM CO If) H £> CJ\ O 00 C* O 

01 "t CO N N C\ H H * Tj-M l^CO'*- 

i-i • A ■ 

■^-OOONONi-iO^coOOOO 
co 

. rt .w - i i o3 w> c « + 1 



and kangankunde, malawi 205 

Table 3 

MW71 MW135 MW71 and MWi35are 
%Na 2 0-70 0-63 feldspathic breccias 

%K 2 i5 - 8g 15-90 from Chilwa Island 

rocks developed in situ from basement schists and syenites points to a very liberal 
redistribution of elements. The general chemical changes which have been recog- 
nized to be typical of fenitization are nicely shown by the analyses of Kangankunde 
fenites and the associated metasomatic veins. It must be stressed that the veins 
(analyses 10 and 12, Table 1) represent in situ alteration of rocks 9 and 11 (Table 1). 
The changes are consistent and reveal weight percentage decreases in Si0 2 , FeO, 
CaO and MgO and increases in Ti0 2 , Fe 2 3 , MnO, Na 2 0, K 2 0, P 2 5 and water. 
Although these changes are sometimes small, their constancy is telling. Of out- 
standing significance is the large increase in the oxidation ratio, and the increase of 
the potash to soda ratio. These two increases are particularly characteristic of the 
change from the syenitic fenites to the ultra-potassic rocks, the feldspathic breccias 
of Garson. 

It has been found that the triangle Oz-Ne-Ks is a particularly useful one for 
following the changes in the felsic constituents during fenitization, anorthite being 
insignificant in most fenites. In Text-figs. 5, 6, 7 and 8, plots of fenites from Chilwa 
Island and Kangankunde, Alno, and various other fenite localities are presented. 
The Alno results (Text-fig. 5) are particularly revealing; the mobilized ultra-fenites 
of von Eckermann (1948, p. 43) have been excluded. The tie-lines link rocks 
collected along recognizable horizons in the basement by von Eckermann to show the 
progressive fenitization of rocks which it can reasonably be assumed were once 
chemically similar. Two stages of fenitization are brought out by the diagram; the 
initial change from saturated country rocks and quartz syenites to saturated syenitic 
fenites lying close to the Ab-Or join, then the persistent increase in the K 2 0/Na 2 
ratio moving the plots towards the composition of orthoclase. If the Chilwa Island 
fenites are now considered (Text-fig. 6) they show a somewhat similar pattern 
varying from oversaturated quartz fenites to syenitic fenites plotting on the Ab-Or 
join. This change is less extreme than the Alno one because unfenitized rocks are 
not found on Chilwa Island, and the older syenites, when unfenitized, are poorer in 
silica than the average schist or gneiss. The alteration of syenitic fenites to the 
feldspathic breccias is, however, more extreme than the potash enrichment at Alno, 
as shown by the dashed line on Text-fig. 6 : a reflection of the almost total removal 
of sodium from the system. The aegirine-feldspar vein from Chilwa Island plots, 
rather surprisingly, close to the orthoclase point indicating the co-existence of a 
potassium-rich feldspar with the soda-pyroxene. The Kangankunde analyses are 
joined by tie-lines which show the trend of alteration along the veins. Analyses 9 
and 10 show only a slight bulk change but what change there is is similar to that at 
Chilwa Island. Analyses n and 12 represent a much more fundamental alteration 
from a quartz-rich rock to one which is only slightly oversaturated. These Kangan- 
kunde results must, however, be considered in the light of the fact that the changes 
in individual oxides reflect a trend which is more consistent with the change to a 



206 



FENITIZATION AT CHILWA ISLAND 



feldspathic breccia than to a syenitic fenite, while the microscopic evidence suggests 
that the veins are partly zoned from syenitic fenites through to potash-feldspar-ore 
rocks. Analysis 12, therefore, represents an admixture of syenitic fenite with potash- 
feldspar-ore rock, and reveals a trend which would be expected from such a hybrid. 




Fig. 5. Plot of fenites and associated country rocks of the Alno Complex in the system 
Qz— Ne— Ks. Data taken from von Eckermann (1948, analyses 1-20). Tie lines join 
specimens collected along the strike. O = country rocks ; ^ = quartz fenite ; • = sye- 
nitic fenites; □ = nepheline-bearing fenites. 



The chemical changes during fenitization in terms of the system Qz-Ne-Ks for a 
number of complexes, including Chilwa Island and Alno, are given as Text-fig. 7. 
The lines represent trends shown by two or more analyses, and the dashed lines 
indicate the phase of potassium enrichment at Alno and Chilwa. The majority of 
published fenite analyses are plotted on Text-fig. 8. Text-figs. 7 and 8 indicate an 
initial impoverishment in silica, then a certain amount of spread in the syenitic 
fenites between soda feldspar-bearing and potash feldspar-bearing rocks, the latter 
represented by the distinct trend towards the ultra-potassic rocks. It is noticeable 
that there is a slight concentration of syenitic fenite plots around the 35 % potash- 
feldspar position on the Ab-Or join corresponding to the experimentally determined 
minimum for the alkali feldspars (Schairer, 1950). The trend lines of Text-fig. 7, 
with but two exceptions, radiate from the quartz apex indicating that there is little 



AND KANGANKUNDE, MALAWI 



207 



change in the bulk composition of the feldspars during fenitization, the total increase 
of sodium being partly held in the pyroxene. On Text-fig. 8 a few nepheline fenites 
are shown. 

That there is an increase in alkalis during fenitization has long been realized and 
this is brought out well in Text-fig. 9 in which the trend of alkali enrichment, con- 
comitant with a decrease of silica during the fenitization, is almost perpendicular to 




Fig. 6. Plot of Chilwa Island and Kangankunde fenites in the system Qz-Ne-Ks. The 
analyses of Kangankunde quartz fenites and the associated cross-cutting fenite veins, 
are joined by tie lines. A = quartz fenites; • = syenitic fenites; ■ = feldspathic 
breccia; X = aegirine-potash-feldspar fenite vein. Dashed arrow indicates path of 
f eldspathization . 



the alkali-silica relationship for normal rock series. The Chilwa Island and Kangan- 
kunde results conform to the trend, the latter being joined by tie lines which show the 
extreme alkali enrichment and silica impoverishment of analysis 12 in relation to ir. 
At the left hand end of the normal rock series on Text-fig. 9 are plotted a gabbro and 
a basalt and their fenitized equivalents, taken from Mathias (1956, Table 4). These 
plots show that the general tendency of alkali enrichment holds for basic as well as 
acid fenitized rocks, although the change in silica is now insignificant. 

As well as an inverse relationship between silica and alkalis during fenitization 
there is a similar relationship between silica and iron, magnesia, and lime which is 



208 



FENITIZATION AT CHILWA ISLAND 



brought out in Text-fig. io, and applies at Chilwa Island and Kangankunde. There 
is an increase of all three elements, although in general the increase in iron is the most 
substantial. Because of the abnormal increase of alkalis during fenitization the 
increase of total iron, magnesia, and lime is less than for a normal rock series showing 
a similar decrease of silica, and this is shown in Text-fig. io in which the normal trend 
is indicated by the dashed line. The low values for the feldspathic breccias are 
apparent. 




Fig. 7. The trend of fenitization at various localities plotted in the system Qz-Ne-Ks. 
The dashed lines trace the trend of potash enrichment at Chilwa Island and Alno. 
A = Chilwa Island (8 analyses); B = Alno (von Eckermann, 1948 — 20 analyses); 
C = Oldoinyo Dili (McKie, 1966 — 12 analyses); D = Norra Karr (Adamson, 1944 — 2 
analyses); E = Chishanya (Swift, 1952 — 3 analyses); F = Fen (Saether, 1957 — 3 
analyses); G = Iivaara (Lehijarvi, i960 — 2 analyses); H = Spitzkop (Strauss & Truter, 
1951 — 10 analyses); I = Dorowa (Johnson, 1961 — 4 analyses). 



It has been shown by several workers that there is an increase in the oxidation 
ratio (Fe 3+ /Fe 3+ + Fe 2+ ) during fenitization (von Eckermann 1948, p. 32; Verwoerd, 
1966, p. 136; McKie, 1966, p. 278), ascribed by von Eckermann to the oxidizing 
action of C0 2 emanating from the carbonatite. The Chilwa Island and Kangan- 
kunde rocks conform very much to this pattern (Text-fig. 11) as shown bjr the 
increasing oxidation ratio of the Chilwa Island fenites, the higher value still of the 
fenite vein, and the characteristic total absence of ferrous iron in the feldspathic 
breccias. The high oxidation ratios revealed by the analyses of the Kangankunde 



AND KANGANKUNDE, MALAWI 



209 



vein rocks indicates that they have been subjected to chemical changes more akin 
to those which promote the development of the feldspathic breccias than the syenitic 
fenites. 

DISCUSSION 

Before considering some of the genetic aspects of fenitization in the Chilwa Series, 
some clarification of the nomenclature of fenites, as used in this account, is needed. 
Three quite distinct major processes, and a number of minor ones, are commonly 




Fig. 8. Composite plot of fenite analyses in the system Qz-Ne-Ks. O = country rocks; 
^ = quartz fenites; • = syenitic fenites; ■ = ultra-potassic fenites; □ = nepheline- 
bearing fenites. 



referred to the omnibus term fenitization, and certain aspects pertaining to the 
nature of these three branches of fenitization can be illustrated by Text-fig. 12. 
This diagram has been constructed from the data presented in Text-fig. 8, which 
incorporates a high proportion of the chemical data on fenites published to date. 
On Text-fig. 12, three metasomatic processes are considered: fenitization proper; 
feldspathization (as used by Garson & Campbell Smith, 1958), and nephelinization. 
It seems reasonable to distinguish these three processes because each one is chemi- 
cally and mineralogically distinctive, and, as will be discussed shortly, they are often 
associated with different groups of magmatic rocks : — 

Fenitization is promoted by syenites, nepheline syenites, ijolites, carbonatites. 

Nephelinization is promoted by nepheline syenites, ijolites, carbonatites(P). 



2IO 



FENITIZATION AT CHILWA ISLAND 



Feldspathization is promoted by carbonatites. 

Chemically the three processes, as they apply to granitic rocks (granites, gneisses 
and schists), are characterized as follows: 

Fenitization = decrease of Si (from oversaturated to saturated rocks) increase 

of Na, K, Fe, Mg, Ca, Ti, and P and gradual increase of oxidation 

ratio. 

Nephelinization = decrease of Si (from saturated to undersaturated rocks) — 

chemical data then uncertain, but probable increase of 

alkalies. 

Feldspathization = sharp increase of oxidation ratio towards ioo; radical 

increase of K 2 0: Na 2 ratio; decrease of Fe, Mg, Ca. 



12 



10 



Na + K 



*FG 






V- 




• • 



• □ 



•NT. 



X ■- A 



"°B 



<<^i 




28 



30 



32 



34 



36 



38 



40 



42 



Si 



Fig. g. The variation of alkalis (Na + K) and silica during fenitization, based on a cell 
of ioo oxygens. The main trend is indicated on the small inset diagram. The dashed 
line represents the variation of alkalis and silica for a normal rock series. Symbols as in 
fig. 8. B = basalt; G = gabbro; FB = fenitized basalt; FG = fenitized gabbro: from 
Messum Complex (Mathias, 1956). Tie lines join plots of Kangankunde veined fenites. 



AND KANGANKUNDE, MALAWI 



211 



20 



18 



16 



14 



12 



10 



fcFG 



PG 
\ 



fc FB 

o 



Ca + Mg + 
total Fe 



8 



o 

B \ 






\ 
\ 



\ 



<0 



\ 
\ 
\ 



\ 



\ 



\ 



\ 



-v. 



\ 



*•• 



\ 



o« 



• «A 



\ 
o 
o\ 

\ 

° \ 

A O 

I 

i 
AAA 



I 



I 



1 



<*£ oo 
o\ 

\ 



28 



30 



32 



38 



40 



34 36 

Si 

Fig. io. The variation of (Ca + Mg + total Fe) against silica during fenitization, based on 
a cell of ioo oxygens. The principal trend is indicated on the small inset diagram. The 
dashed line indicates the trend for normal rock series. Symbols are as on fig. 8. 



42 



212 



FENITIZATION AT CHILWA ISLAND 

100 r 



80 



60 
100 Fe 3 + 

Fe 3+ +Fe 2+ 

40 



20 




_L 



32 



34 



36 
Si 



38 



40 



Fig. ii. Variation of the oxidation ratio (ioo Fe 3+ /Fe 3+ + Fe 2+ ) with the change in silica 
content for Chilwa Island and Kangankunde (Table i). # = Chilwa Island fenites; 
± = Kangankunde fenites; X = aegirine-potash-feldspar fenite vein, Chilwa Island; 
■ = feldspathic breccia Chilwa Island and two feldspathic breccias from the Tundulu 
Complex (Garson, 1962, p. 79). 



These chemical changes reflect the following mineralogical changes — 

Fenitization = loss of quartz, growth of sodic pyroxene and amphibole, 
increase of alkali feldspar, resulting in perthitic syenite with 
alkaline mafics (i.e. aegirine, sodic amphibole). 
Nephelinization = growth of nepheline at the expense of feldspar, with the 

production of a nepheline syenite. 
Feldspathization = breakdown of sodic pyroxene and amphibole, alteration of 

alkali feldspar to potash-feldspar. Production of an ultra- 
potassic syenite. 

The three processes are easily distinguished but many localities have been 
described, including Chilwa Island and Kangankunde, at which two or even three of 
these changes are contiguous, that is fenitization grading into nephelinization and 
fenitization grading into feldspathization, while fenitization and feldspathization may 
occur independently. A particular difficulty which arises in considering these three 
metasomatic processes however, is in deciding how much of the fenitization and 
nephelinization is caused by the carbonatite, a problem that has been discussed by 
King & Sutherland, (i960, p. 714). Because of the lack of large masses of silicate 
rocks in the Chilwa Island and Kangankunde complexes it would at first appear 
that all the metasomatic effects must be attributed to the carbonatite, and this is 
Garson's conclusion with regard to the Kangankunde complex (1965a, p. 123). 



AND KANGANKUNDE, MALAWI 
Qz 



213 



Ne 




ULTRA-POTASSIC 
ROCKS 

Or 



Feldspathisat ion 



Nephelinisation 



Na.KAI.Si.O,. 
3 4 4 16 



Fig. 12. Generalized diagram to show the principal chemical changes during fenitization, 
feldspathization and nephelinization, as they can be represented in the system Qz-Ne-Ks. 



Dawson in his discussion of this very problem (1964, p. 108) cites Chilwa Island as a 
locality at which the fenitization is indubitably attributable to the carbonatite. 
Although this may be correct there is every reason to suppose that the Chilwa and 
possibly the Kangankunde complexes are sub- volcanic. As such, a pipe, now occu- 
pied by carbonatite, almost certainly was once a volcanic conduit facilitating the 
passage of strongly alkaline, undersaturated magmas to the surface. These magmas, 
which were probably phonolitic and may be represented by the Lupata Gorge 
volcanics occurring some 140 miles to the south-west (Dixey & Campbell Smith, 
1929), were probably capable of fenitizing the walls of the pipe. This certainly 
applied to many of the circular or oval, sub- volcanic carbonatite complexes, which at 
their present erosion level reveal little, or no, silicate rock, but which have a fenite 
aureole. The metasomatic changes at Chilwa Island, Kangankunde and other 
Malawi centres which are undoubtedly attributable to the carbonatite are the felds- 
pathization, which Garson has shown by his mapping at Kangankunde forms 
collars around the carbonatite pipes (Garson, 1965a), and the process of phlogopitiza- 
tion at Kangankunde. In an attempt to decide this issue the contacts of carbona- 
tite bodies where there was definitely no influence from silicate magma must be 



2i 4 FENITIZATION AT CHILWA ISLAND 

considered. One such carbonatite is that at Songwe Scarp, described by Brown 
(1964, p. 223). The main contact effect according to Brown is a pronounced potash 
feldspathization of the adjacent schists; a marked increase in the oxidation ratio is 
also apparent from the analyses (op. cit. Table 3). Similarly the Rufunsa carbona- 
tites (Bailey, i960) only potash-feldspathize the country rocks, there is no sodic 
pyroxene or amphibole development. Fenitization of granites by beforsite dykes 
at the confluence of the Elands and Crocodile Rivers, Transvaal has been described 
by Fockema (1953, p. 155) but in fact the fenitized rocks consist of feldspar, quartz, 
chlorite and hydrated iron oxides, and analyses reveal a very high potash to soda 
ratio, suggesting feldspathization rather than fenitization. In contrast to these 
localities where an alteration akin to feldspathization rather than fenitization 
prevails, von Eckermann has found a definite increase of fenitization around sovite 
dykes at Alno (1948, PI. 59). The same feature is apparent at Tundulu (M. S. Garson; 
personal communication). These contrasting examples suggest either that there are 
a range of carbonatite magmas having strongly contrasting metasomatizing capabili- 
ties, perhaps related by an evolutionary sequence of changes in alkali content, or that 
a phenomenon of differential zonation related to contrasting rates of loss of Na and K 
from the magma occurs, or yet again that the fenitizing ability of carbonatite is 
dictated by the prevailing physico-chemical conditions. 

It has been realized for some time that fenitization is not restricted to carbonatite 
complexes, and is to be found around nepheline syenite and ijolite bodies. Even 
saturated syenites are capable of fenitization as evidenced by the introduction of 
alkalis into Moine Schists around the Cnoc nan Cuilean intrusion, Sutherland, 
Scotland, resulting in the growth of aegirine-augite, as described by King (1943, 
p. 147). Garson has described fenitization effects around syenite bodies in the South 
Mlanje Area of Malawi (1963, p. 12) with the development of aegirine-augite in the 
altered country rocks. Syenites are only capable of promoting fenitization, that is 
alteration towards saturated rocks — syenitic fenites, but the nepheline syenites and 
ijolites are capable of desilicating the enclosing rock envelope — nephelinization. 
These changes are brought out in Text-fig. 12 and it is obvious that in terms of 
Qz-Ne-Ks the metasomatic changes are such as to make the altered rock over to a 
composition near to that of the active intrusive rock, so that the composition of the 
original material is not particularly significant. 

It seems very unlikely that nephelinization can be caused by carbonatite, although 
at Alno von Eckermann interprets the undersaturated rocks as hybrids produced by 
the interaction of carbonatite and other rocks (1948). The nephelinization at 
Tundulu has been shown by Garson to be related to the main ring-dyke of foyaite 
(1962, p. 61), while at Dorowa the nephelinized fenite (pulaskitic fenite) surrounds a 
body of foyaite and ijolite (Johnson 1961, p. 106). The inclusion of blocks of felds- 
pathic breccia in carbonatite at Chilwa Island and Kangankunde indicates that the 
carbonatite was incapable of nephelinization, at least at this level. Inclusions of 
potash feldspar have been described from many carbonatites, indicating its stability 
in some carbonatite magmas. 

The process of phlogopitization at Kangankunde is shown by Garson & Campbell 
Smith (1965a, p. 44) to be a hydrothermal process involving the addition of MgO 



AND KANGANKUNDE, MALAWI 215 

and water from ankeritic carbonatite, to potash feldspar. This process is not 
apparent at Chilwa Island, probably because the ankeritic carbonatite is insulated 
from the fenites by the earlier sovite. A similar process took place at Fen as 
can be seen in rocks on the W. side of the entrance to Hydro's quarry where blocks of 
fenite, enveloped in carbonatite, have reaction rims of phlogopite. The intermediate 
process of potash-feldspathization is not, however, manifest in the Fen rocks. 

Before considering the relative importance of fenitization, nephelinization and 
feldspathization further it is appropriate to look at the chemical evidence from the 
Oldoinyo Lengai carbonatite lava (Dawson 1962, 1966 ; DuBois et al., 1963). Analyses 
of this lava give K 2 = 6-5-7-6% and Na 2 = 29-0-30-0% (Dawson, 1966, Table 
IV). Analyses of plutonic carbonatite rock show an average total alkali content of 
less than 2%, with the notable feature that K 2 is much greater than Na 2 (Hein- 
rich, 1966, Table 8-2). It seems very probable that most carbonatite magmas were 
at least as well endowed with alkalis as the Oldoinyo Lengai lava but that these alkalis 
were lost to the aureole during emplacement. The Oldoinyo Lengai carbonatite 
obviously has considerably more soda to contribute to the country rocks than potash, 
so that the highly potassic metasomatism associated with the carbonatites of Chilwa 
Island, Kangankunde and other Malawi carbonatite complexes, Songwe and Rufunsa 
presents a problem. Bailey suggests (1966, p. 150) that at Rufunsa the lack of soda 
may be an intrinsic property of the magma, or that at low temperatures soda may 
be unreactive with the wall rocks and hence is expelled at the surface in hot brines 
and as soda-ash. In considering the Malawi complexes in which fenitization and 
feldspathization occur two possibilities present themselves. 

(a) An alkali-rich carbonatite was emplaced and the soda was lost preferentially, 
fenitizing the country rocks; the larger potash ions were then expelled and 
fixed in an inner zone of potash-rich rocks. That is differential zonation. 

(b) The fenites were formed through the agency of silicate magmas in a central 
volcanic pipe, and the feldspathization was then superimposed on these by the 
later emplacement of the carbonatite. The potash metasomatism generated by 
the carbonate could be due either to the fact that potash was the only alkali 
available in the carbonatite, or that soda was held back and then expelled at 
the surface, as suggested by Bailey. 

If the second variation is accepted it suggests that there are a range of carbonatite 
types; those, possibly primary ones, containing abundant soda and potash, and a 
series derived from these of potash-rich and soda-rich carbonatites. i.e. 



xS 



P carbonatite rock + soda-ash 



H H 



carbonatite rock 5 5 and Na-bearing brines 

K< 

K< 



►K § I 



(K) carbonatite | or (Na) carbonatite 

Na< -f >Na < K< f >K 

Na< — I >Na S K< I »K 

(Na + K) carbonatite £ (Na + K) carbonatites 



= 3 

< X 

H f- 

O < 



CO 



216 FENITIZATION AT CHILWA ISLAND 

Presumably alkalis would be lost continuously within the higher levels of the crust 
and the " metasomatizing power " and character of the metasomatism would be 
determined by crustal level, proportion of the alkalis present, and the contrasting 
physico-chemical properties of the migrating potassium and sodium ions. Some 
carbonatite bodies have had very little effect on their immediate surroundings, so it is 
probable that these represent the last passive carbonatite fraction after the alkalis 
have been lost. 

It is apparent that the Chilwa Island and Kangankunde fenitized and felds- 
pathized rocks may have been generated by the differential zonation around an 
alkali carbonatite of Oldoinyo Lengai type, but that the mechanism whereby the 
fenitization is caused by a silicate magma and the ultra-potassic rocks by carbonatite 
is also quite feasible. There does not at the present time appear to be any reliable 
criterion to distinguish the two mechanisms, as they have applied in the Chilwa 
Province. 

The microscopic evidence of Chilwa Island and Kangankunde fenites provides an 
abundance of evidence as to the actual mechanics of the fenitization process. 
Element migration was facilitated by veins, grain boundaries and cracks and cleavages 
in the minerals as shown by the distribution of the turbidity in the feldspars and 
the patterns of sodic pyroxene and amphibole growth. It would appear that 
migration took place principally in a fluid phase probably rich in water and C0 2 and 
that silica and alkalis were the principal migrants, though the threads of ore along 
many fenite veins attest to the mobility of iron. The metasomatic movements 
appear to have been on three levels : a short range movement within single grains, or 
between adjacent grains; a redistribution of material within the aureole as a whole, 
and a larger scale overall removal from, or addition to, the aureole. Small scale 
movement is exemplified, for instance, by the growth of aegirine at the expense of 
quartz with the soda being derived from adjacent plagioclases. Migration within 
the aureole is suggested by the impoverishment of the feldspathic breccias in iron, 
magnesium and sodium, while these elements accumulated in the normal fenitized 
rocks. The largest scale of movement involves the wholesale loss of silica from the 
system and the enrichment in alkalis ; there has probably been an overall increase 
also in phosphorus and titanium. The silica loss provides the greatest problem with 
regard to its ultimate destination. Was it driven outwards, inwards or upwards? 
Von Eckerman suggests that at Alno the silica migrated inwards into the carbonatite 
(von Eckermann, 1948). A combination of inward and upward movement seems 
the most likely, the silica perhaps being carried up a central volcanic pipe incorporated 
in silicate magma, or carried in solution to the surface in brines. The late silicifica- 
tion of the feldspathic breccias suggests that this means of egress from the system was 
closed when the carbonatite was emplaced. 

The mobilization of fenites has been cited from several localities, for instance 
Alno (von Eckermann, 1948, p. 27), and Semarule (King, 1955), and of mobilized 
ultra-potassic rocks from Chilwa Island (Garson & Campbell Smith, 1958, p. 29), 
Tundulu (Garson, 1962, p. 74) and Toror (Sutherland, 1965). Most of the intrusive 
silicate rocks at Alno, nepheline syenites and ijolites, are ascribed to the mobilization 
of " ultrafenites " by von Eckermann. King & Sutherland (i960, p. 719) have 



AND KANGANKUNDE, MALAWI 217 

evolved a scheme based on observations of East African alkaline rocks and carbo- 
natites according to which mobilized fenites were emplaced as nepheline syenites 
and syenites, with equivalent extrusive phases of phonolite and trachyte. The 
Chilwa Alkaline Province is not restricted to the suite of carbonatites and closely 
associated rocks but also includes substantial plutons of syenite, Mlanje and Zomba 
Mountains, and nepheline syenite, Chikala, Chaone and Mongolowe Hills (Bloomfield, 
1965, p. 96; Garson, 1963), which are closely associated in space and time with the 
carbonatites. The only extrusive rocks found in the main area of the Chilwa 
Province are small patches of thermally metamorphosed basic rocks on Chikala Hill 
(Stillman & Cox, i960, p. 102). A chemical analysis of a porphyritic variety proved 
to be similar to the average nepheline tephrite (op. cit. Table 1). Some 140 miles to 
the south-west of Lake Chilwa a volcanic succession is exposed along the Lupata 
Gorge of the Zambezi, consisting of rhyolites, phonolites, kenytes, blairmorites 
(phonolites containing primary analcite) and tuffs, and these are intruded by nephe- 
line syenite, microfoyaite and tinguaite (Dixey & Campbell Smith, 1929). This 
succession is cut by the Muambe carbonatite (Dias, 1961). The Muambe Complex 
belongs to the Chilwa Province and the alkaline lavas are very probably an extrusive 
phase consanguineous with the Chilwa Series. It is suggested that the relationship 
between the carbonatites, the nepheline syenites, syenites and volcanics may be as 
follows : 



PHONOLITES OR 



PHONOLITES AND 



CARBONATITE 




NEPHELINITES TRACHYTES 

f t 

PRIMARY NEPHELINITIC f,„ih,a,ion FENITES AND -nobiliwlion SYENITES AND fatftafen FENITES AND 

PHONOLITIC MAGMA nepM.n.so.ion NEPHELINE FENITES NEPHELINE SYENITES n.ph.lWsot.on NEPHELINE FENITES 

FENITES moblliiQlion ^ SYENITES 

ULTRA- POTASSIC r "° b ''"°''° n ) POTASH -TRACHYTES 

ROCKS 

ACKNOWLEDGEMENTS 

The field work was carried out and the laboratory studies commenced during the 
tenure by the writer of a N.E.R.C. Research Fellowship at Bedford College, which is 
gratefully acknowledged. The Director of the Malawi Geological Survey generously 
helped with camping equipment, labour and transport during field work, while 
Dr. M. Garson introduced the writer to the geology of Chilwa Island and Kangankunde 
and afforded much other help. The subject of this paper was originally suggested by 
Professor B. C. King to whom the writer is grateful for many discussions on the 
problems of fenitization, and for a critical reading of the manuscript. The chemical 
analyses were made by A. J. Easton. 

REFERENCES 

Adamson, O. J. 1944- The petrology of the Norra Karr District. Geol. For. Stockh., 66: 

113-255- 
Bailey, D. K. i960. Carbonatites of the Rufunsa valley, Feira district. Bull. Dep. geol. 

Surv. Nth. Rhod. 5. 



218 FENITIZATION AT CHILWA ISLAND 

Bailey, D. K. 1966. Carbonatite volcanoes and shallow intrusions in Zambia. In Carbonatites 

(editors Tuttle, O. F. and Gittins, J.), Wiley and Sons. 
Bloomfield, K. 1965. The geology of the Zomba area. Bull. geol. Surv. Malawi. 16. 
Brown, P. E. 1964. The Songwe scarp carbonatite and associated feldspathization in the 

Mbeya Range, Tanganyika. Q. Jl. geol. Soc. Lond. 120 : 223-240. 
Dawson, J. B. 1962. Sodium carbonate lavas from Oldoinyo Lengai, Tanganyika. Nature, 

Lond. 195 : 1075-1076. 

1964. Reactivity of the cations in carbonate magmas. Proc. geol. Ass. Can. 15 : 103-113. 

1966. Oldoinyo Lengai — an active volcano with sodium carbonatite lava flows. In 

Carbonatites (editors Tuttle, O. F. and Gittins, J.), Wiley and Sons. 
Deer, W. A., Howie, R. A. and Zussman, J. 1962. Rock-forming minerals, vol. 2. Longmans, 

London. 
Dias, M. de B. 1961. Geologia do Monte Muambe. Bolm. Servs Ind. Minas Geol., Lourenco 

Marq. 27 : 37-64. 
Dixey, F. & Smith, W. Campbell. 1929. The rocks of the Lupata Gorge and the north side 

of the Lower Zambezi. Geol. Mag. 66 : 241-259. 
Dixey, F., Smith, W. Campbell, & Bisset, C. B. 1937. (revised 1955). The Chilwa Series 

of Southern Nyasaland. Bull. geol. Surv. Nyasald. 5. 
Dubois, C. G. B., Furst, J., Guest, N. J. & Jennings, D. J. 1963. Fresh natro carbonatite 

lava from Oldoinyo L'Engai. Nature, Lond. 197 : 445-446. 
v. Eckermann, H. 1948. The alkaline district of Alno Island. Sver. geol. Unders. Ser. Ca. 

No. 36. 
Fockema, R. A. P. 1953. The geology of the area around the confluence of the Elands and 

Crocodile Rivers. Trans. Proc. geol. Soc. S. Afr. 55 : 155-174. 
Garson, M. S. 1962. The Tundulu carbonatite ring-complex in Southern Nyasaland. Mem. 

geol. Surv. Dep. Nyasald. 2. 

1963. The geology of the South Mlanje area. Rec. geol. Surv. Nyasald. 5 : 5-16. 

1965a. Carbonatite and agglomeratic vents in the western Shire Valley. Mem. geol. Surv. 

Dep. Malawi. 3. 

1965&. Carbonatites in Southern Malawi. Bull. geol. Surv. Nyasald. 15. 

1966. Carbonatites in Malawi. In Carbonatites (editors Tuttle, O. F. and Gittins, J.), 

Wiley and Sons. 

& Smith, W. Campbell. 1958. Chilwa Island. Mem. geol. Surv. Dep. Nyasald. 1. 

Heinrich, E. Wm. 1966. Geology of carbonatites. Rand McNally and Co. 

Johnson, R. L. 1961 . The geology of the Dorowa and Shawa carbonatite complexes, Southern 

Rhodesia. Trans. Proc. geol. Soc. S. Afr. 64 : 101-146. 
King, B. C. 1943. The Cnoc nan Cuilean area of the Ben Loyal igneous complex. Q. Jl. 

geol. Soc Lond. 98 : 147-185. 
1955- Syenitisation de granites a Semarule, pres de Molepolole, protectorat du Bechuana- 

land. Colloques int. Cent. natn. Rech. scient. 68, Paris, 1-16. 
— ■ — & Sutherland, D. S. i960. Alkaline rocks of eastern and southern Africa. Sci. Progr., 

Lond. 48 : 298-321, 504-524, 709-720. 
Lehijarvi, M. i960. The alkaline district of Iivaara, Kuusamo, Finland. Bull. Commn. 

geol. Finl. No. 185. 
Mathias, M. 1956. The petrology of the Messum igneous complex, South-west Africa. Trans. 

Proc. geol. Soc. S. Afr. 59 : 23-58. 
McKie, D. 1966. Fenitization. In Carbonatites (editors Tuttle, O. F. and Gittins, J.), 

Wiley and Sons. 
Saether, E. 1957. The alkaline rock province of the Fen area in southern Norway. K. 

norske Vidensk. Selsk. Skr. NR 1. 
Schairer, J. F. 1950. The alkali-feldspar join in the system NaAlSi0 4 -KAlSi0 4 -Si0 2 . 

/. Geol. 58 : 512-517. 
Stillman, C. J. & Cox, K. G. i960. The Chikala Hill syenite-complex of Southern Nyasaland. 

Trans. Proc. geol. Soc. S. Afr. 63 : 99-117. 



AND KANGANKUNDE, MALAWI 219 

Strauss, E. A. & Truter, F. C. 1951. The alkali complex of Spitskop, Sekukuniland, 

Eastern Transvaal. Trans. Proc. geol. Soc. S. Afr. 53 : 81-130. 
Sutherland, D. S. 1965. Potash-trachytes and ultra-potassic rocks associated with the 

carbonatite complex of the Toror Hills, Uganda. Mineralog. Mag. 35 : 363-378. 
Swift, W. H. 1952. The geology of Chishanya, Buhera District, Southern Rhodesia. Trans. 

Edinb. geol. Soc. 15 : 346-359. 
Verwoerd, W. J. 1966. South African carbonatites and their probable mode of origin. 

Annale Univ. Stellenbosch. 41, Ser. A, No. 2. 



A. R. Woolley, B.Sc, Ph.D., F.G.S. 

Department of Mineralogy 

British Museum (Natural History) 

London, S.W.7 



PLATE 9 

A. Electron probe traverses across veins of fenitization in plagioclases of a Kangankunde 
fenite (B.M. 1968, P 37, 267). Along the veins the plagioclase twinning is destroyed and a new- 
turbid feldspar develops. The probe traverses, the paths of which are indicated by the ruled 
lines on the photomicrographs, show that sodium is concentrated in the unaltered plagioclase 
and in the central part of the veins, which is occupied by aegirine. The new turbid feldspar is 
poor in sodium and rich in potassium. The scale bar between the photomicrographs is 0-05 cm. 
long. 

B and C. Kangankunde fenites cut by feldspathization veins {(B = B.M. 1968, P 37, 284; 
C = B.M. 1968, P 37, 299)}. The central black, thin thread in each case is predominantly 
limonite while the bordering feldspathized area is pink in colour, in contrast to the black and 
grey mottled fenite. The scale lines are 1 cm. long. The veined and unveined parts of these 
rocks were mechanically separated and chemically analysed (Table 1 , analyses 9-1 2) . 



Bull. Br. Mus. nat. Hist. (Miner.) 2, 4 



PLATE 9 





PRINTED IN GREAT BRITAIN 
BY ADLARD & SON LIMITED 
BARTHOLOMEW PRESS, DORKING